[Matt Devo]

Where are you copying these files from? Do they have proper ownership and permissions? Do they have proper Unix formatting or were they created on windows?

Via SMB. Thats why its so odd. AppleTV, Samsung Tablet, RasberryPi, and Win7 all use the same exact files but the CB is having issues with them. I guess I will try and reinstall openELEC and see if it works.

under System->OpenELEC->Network, ensure that option 'Wait for network before starting XBMC' is set/selected

[igjunk]

Hi Matt,

In the release notes of OpenELEC 4.2.1 mentioned a fix for Intel GPU stuttering. Do you know if it's related to the VA-API bug which requires software processing?

[Matt Devo]

Hi Matt,

In the release notes of OpenELEC 4.2.1 mentioned a fix for Intel GPU stuttering. Do you know if it's related to the VA-API bug which requires software processing?

I believe this only related to 25/50p sources, but not 100% sure. You shouldn't be using software decoding though, the bug requiring that was fixed a long time ago. Just use the recommended settings in the wiki.

[fritsch]

In fact the GPU driver OE 4.0 used was bad. This driver could not correctly sync to the VideoClock. One did not see skips or drops but the image stuttered never the less, utter crap. Which VAAPI bug needs software processing? You speak of the VC-1 Interlaced files? That's not really a bug - intel just did not implement support for it and does not plan to do so.

There are far worse bugs which intel does not like to fix currently. For example that one: https://bugs.freedesktop.org/show_bug.cgi?id=80229

[Matt Devo]

In fact the GPU driver OE 4.0 used was bad. This driver could not correctly sync to the VideoClock. One did not see skips or drops but the image stuttered never the less, utter crap. Which VAAPI bug needs software processing? You speak of the VC-1 Interlaced files? That's not really a bug - intel just did not implement support for it and does not plan to do so.

There are far worse bugs which intel does not like to fix currently. For example that one: https://bugs.freedesktop.org/show_bug.cgi?id=80229

think he was referring to this one, for which the only workaround (before the driver fix) was to disable VAAPI hardware acceleration

is the driver in 4.2.1 better in that regard?

[igjunk]

Sorry, I should have been more clear. I was asking whether the need for "Use SW filter for VAAPI" has been fixed with this version.

[Matt Devo]

Sorry, I should have been more clear. I was asking whether the need for "Use SW filter for VAAPI" has been fixed with this version.

I don't believe so

[fritsch]

This bug has nothing to do with xbmc. It's a kernel bug the intel devs don't fix since several months. I linked it above.

[geeQ]

Hello I have a newbie question, I installed openelec on a hp chromebox how d I update the coreboot? I've tried getting into recovery mode and ssh but no luck, Thanks

Never mind re read thru the wiki found my answer

[lburrowes]

Where are you copying these files from? Do they have proper ownership and permissions? Do they have proper Unix formatting or were they created on windows?

Via SMB. Thats why its so odd. AppleTV, Samsung Tablet, RasberryPi, and Win7 all use the same exact files but the CB is having issues with them. I guess I will try and reinstall openELEC and see if it works.

under System->OpenELEC->Network, ensure that option 'Wait for network before starting XBMC' is set/selected

SUCCESS!!! That is all it was. I never had this option selected on my other installs of OPENELEC but I guess the CB boots so fast it must need it... now no issues at all...

[superdesi]

Matt......thanks very much for your hard work. I just got on the bandwagon and purchased a CB today.

With regards to the bios, how is the custom firmware different from say running the command 'chromeos-firmwareupdate --mode=todev' ?

Thanks in advance.

[Matt Devo]

Matt......thanks very much for your hard work. I just got on the bandwagon and purchased a CB today.

With regards to the bios, how is the custom firmware different from say running the command 'chromeos-firmwareupdate --mode=todev' ?

Thanks in advance.

The stock firmware is based on Google's own coreboot branch (vs the mainline open-source one), has has payloads for both UEFI secure boot (depthcharge - to securely boot ChromeOS) and legacy BIOS boot (SeaBIOS - which is completely broken in the stock firmware, so if you want to run anything other than ChromeOS, you need to at the very least update the SeaBIOS payload, which my script does).

Since Google develops both the firmware and ChromeOS kernel, bugs in the former can be worked around in the latter, but obviously that doesn't apply when running anything other than ChromeOS. The stock firmware is very buggy when it comes to suspend/resume and USB3 port handling. And since Google does not have any mechanism in place for updating any part of the firmware, you're stuck with what ships with the box.

My coreboot build is using the latest mainline/upstream coreboot code (into which I integrated support for the ChromeBox using Google's branch), along with the latest mainline/upstream SeaBIOS code. It also contains additional tweaks/fixes by me, such as improved USB3 support, VMX support, and headless booting. It also removes the UEFI component, since it's not needed, resulting in very speedy boot times (~5s).

[superdesi]

Thanks Matt. Im in the process of running your script.

[Sunflux]

So far so good on my new HP Chromebox!

Performed the reset procedure, removed the write screw, ran your script and got OpenELEC installed... then got all my network shares going and built the library, copied the Confluence skin so I could edit it, created a SAMBA link to the /addons/ directory, and got my custom home screen running.

I've tested the Rosewill RHRC-11002, which includes a small whistle-shaped dongle (model IR605) that fits on the front of mini PCs like the NUC and Chromebox, and as far as I can tell it seems to work fine. It will resume from suspend, although it seems to require a bit longer of a keypress than normal to actually wake up.

Here's a photo:



Note that the Rosewill remote is garbage - tiny keys, bad layout, poor repeat performance. But use the receiver with a universal remote or a "real" MCE remote, and the performance is quite good.

If you plan on using the Chromebox with an IR repeater system using stick-on IR emitters, the dongle IS small enough for the back panel USB ports, squeezed between the HDMI and network cables, for a very clean installation. Further, if you use the top port, it leaves just enough room for a slender USB plug to be installed on the bottom port.

[nickr]

You might want to put your comments on that remote onto the wiki.